Re: [PATCH v4 1/5] sched/fair: Clean up attach_entity_load_avg()

2016-06-06 Thread Peter Zijlstra
On Mon, Jun 06, 2016 at 03:40:18PM +0200, Vincent Guittot wrote: > > What about when tasks are moved between groups on CONFIG_SMP=n ? > > Doesn't this change fail to age the load in that case? > > The load average stuff is only enable for SMP system but it's not used > for UP system. > To be

Re: [PATCH v4 1/5] sched/fair: Clean up attach_entity_load_avg()

2016-06-06 Thread Peter Zijlstra
On Mon, Jun 06, 2016 at 03:40:18PM +0200, Vincent Guittot wrote: > > What about when tasks are moved between groups on CONFIG_SMP=n ? > > Doesn't this change fail to age the load in that case? > > The load average stuff is only enable for SMP system but it's not used > for UP system. > To be

Re: [PATCH v4 1/5] sched/fair: Clean up attach_entity_load_avg()

2016-06-06 Thread Matt Fleming
On Mon, 06 Jun, at 03:40:18PM, Vincent Guittot wrote: > > The load average stuff is only enable for SMP system but it's not used > for UP system. Duh, right. My bad.

Re: [PATCH v4 1/5] sched/fair: Clean up attach_entity_load_avg()

2016-06-06 Thread Matt Fleming
On Mon, 06 Jun, at 03:40:18PM, Vincent Guittot wrote: > > The load average stuff is only enable for SMP system but it's not used > for UP system. Duh, right. My bad.

Re: [PATCH v4 1/5] sched/fair: Clean up attach_entity_load_avg()

2016-06-06 Thread Vincent Guittot
Hi Matt, On 6 June 2016 at 15:30, Matt Fleming wrote: > On Mon, 06 Jun, at 08:20:37AM, Yuyang Du wrote: >> attach_entity_load_avg() is called (indirectly) from: >> >> - switched_to_fair(): switch between classes to fair >> - task_move_group_fair(): move between task

Re: [PATCH v4 1/5] sched/fair: Clean up attach_entity_load_avg()

2016-06-06 Thread Vincent Guittot
Hi Matt, On 6 June 2016 at 15:30, Matt Fleming wrote: > On Mon, 06 Jun, at 08:20:37AM, Yuyang Du wrote: >> attach_entity_load_avg() is called (indirectly) from: >> >> - switched_to_fair(): switch between classes to fair >> - task_move_group_fair(): move between task groups >> -

Re: [PATCH v4 1/5] sched/fair: Clean up attach_entity_load_avg()

2016-06-06 Thread Matt Fleming
On Mon, 06 Jun, at 08:20:37AM, Yuyang Du wrote: > attach_entity_load_avg() is called (indirectly) from: > > - switched_to_fair(): switch between classes to fair > - task_move_group_fair(): move between task groups > - enqueue_entity_load_avg(): enqueue entity > > Only in switched_to_fair() is

Re: [PATCH v4 1/5] sched/fair: Clean up attach_entity_load_avg()

2016-06-06 Thread Matt Fleming
On Mon, 06 Jun, at 08:20:37AM, Yuyang Du wrote: > attach_entity_load_avg() is called (indirectly) from: > > - switched_to_fair(): switch between classes to fair > - task_move_group_fair(): move between task groups > - enqueue_entity_load_avg(): enqueue entity > > Only in switched_to_fair() is

[PATCH v4 1/5] sched/fair: Clean up attach_entity_load_avg()

2016-06-06 Thread Yuyang Du
attach_entity_load_avg() is called (indirectly) from: - switched_to_fair(): switch between classes to fair - task_move_group_fair(): move between task groups - enqueue_entity_load_avg(): enqueue entity Only in switched_to_fair() is it possible that the task's last_update_time is not 0 and

[PATCH v4 1/5] sched/fair: Clean up attach_entity_load_avg()

2016-06-06 Thread Yuyang Du
attach_entity_load_avg() is called (indirectly) from: - switched_to_fair(): switch between classes to fair - task_move_group_fair(): move between task groups - enqueue_entity_load_avg(): enqueue entity Only in switched_to_fair() is it possible that the task's last_update_time is not 0 and